A hallmark of typical amnestic Alzheimer's disease (aAD) is neurofibrillary tau pathology. Braak first described staging of tau pathology in AD, classically evident early in the medial temporal lobe (MTL) implicated in impaired episodic memory, and then spreading to neocortical regions important for language, visuospatial and executive function and to basal ganglia for motor function. However, AD pathological change can also present clinically as a focal neocortical syndrome without amnesia, known as non-amnestic Mild Cognitive Impairment (naMCI). This includes disorders of language (logopenic variant primary progressive aphasia, lvPPA), visuospatial (posterior cortical atrophy, PCA), executive (frontal variant MCI, fvMCI), and motor (corticobasal syndrome due to AD, CBS-AD) function. Rare clinical, imaging, and autopsy studies suggest that tau pathology in naMCI is greater in neocortex than MTL, challenging fundamental assumptions about the MTL origin of cerebral tau pathology in AD. However, staging tau pathology at autopsy in naMCI has not been investigated, and longitudinal in vivo imaging to validate spreading disease in naMCI is very rare and involves small groups of clinical cases using a single imaging modality. These represent major gaps in our knowledge. In Aim 1, we study stages of spreading tau pathology at autopsy in naMCI and aAD with a novel, validated, digital method. We hypothesize that stages of accumulating tau pathology are consistent with neocortical origin and spread in naMCI, and only later spread to MTL, differing from the MTL origin of tau pathology in aAD. Pathologic staging depends on inferences from cross-sectional studies at autopsy, so Aim 2 proposes to validate stages of spreading disease in vivo with novel longitudinal MRI and tau-PET imaging in naMCI and aMCI/ aAD using unique imaging and graph theory approaches. We hypothesize a cortical origin and spread of tau in naMCI, with later accumulation of tau in MTL, reversing the MTL origin of disease in aMCI/ aAD. These findings would challenge assumptions about the MTL origin of tau pathology in AD. We test the novel hypothesis that genetic factors in FTLD-tau bias the anatomic distribution of tau pathology contributing to the atypical spread of tau in naMCI. In Aim 1, we expect that stages of tau pathology in naMCI resemble that seen in primary tauopathies due to FTLD-tau, including non-fluent/agrammatic variant PPA, behavioral variant frontotemporal dementia, and corticobasal degeneration/progressive supranuclear palsy. In Aim 2, we expect that patterns of longitudinal in vivo imaging in naMCI resemble that seen in FTLD-tau. Finally, Aim 3 proposes a targeted study of FTLD- tau-related single nucleotide polymorphism (SNP) risk alleles in naMCI. Based on our preliminary findings, we hypothesize that FTLD-tau-related risk alleles are more common in naMCI than aMCI/ aAD. These findings would challenge long-held assumptions about the pathophysiologic basis for AD pathology, develop validated diagnostic criteria for naMCI to support inclusion in disease-modifying AD treatment trials, lead to validated endpoints for these trials, and provide needed prognostic information for this underserved population.